Time-series machine learning at scale

functime is a powerful Python library for production-ready global forecasting and time-series feature extraction on large panel datasets.

functime also comes with time-series preprocessing (box-cox, differencing etc), cross-validation splitters (expanding and sliding window), and forecast metrics (MASE, SMAPE etc). All optimized as lazy Polars transforms.

Join us on Discord!

Highlights

Fast: Forecast and extract features (e.g. tsfresh, Catch22) across 100,000 time series in seconds on your laptop
Efficient: Embarrassingly parallel feature engineering for time-series using Polars
Battle-tested: Machine learning algorithms that deliver real business impact and win competitions
Exogenous features: supported by every forecaster
Backtesting with expanding window and sliding window splitters
Automated lags and hyperparameter tuning using FLAML

Additional Highlights

functime comes with a specialized LLM agent to analyze, describe, and compare your forecasts. Check out the walkthrough here.

Getting Started

Install functime via the pip package manager.

pip install functime

functime comes with extra options. For example, to install functime with large-language model (LLM) and lightgbm features:

pip install "functime[llm,lgb]"

cat: To use catboost forecaster
xgb: To use xgboost forecaster
lgb: To use lightgbm forecaster
llm: To use the LLM-powered forecast analyst

Forecasting

import polars as pl
from functime.cross_validation import train_test_split
from functime.seasonality import add_fourier_terms
from functime.forecasting import linear_model
from functime.preprocessing import scale
from functime.metrics import mase

# Load commodities price data
y = pl.read_parquet("https://github.com/functime-org/functime/raw/main/data/commodities.parquet")
entity_col, time_col = y.columns[:2]

# Time series split
y_train, y_test = y.pipe(train_test_split(test_size=3))

# Fit-predict
forecaster = linear_model(freq="1mo", lags=24)
forecaster.fit(y=y_train)
y_pred = forecaster.predict(fh=3)

# functime ❤️ functional design
# fit-predict in a single line
y_pred = linear_model(freq="1mo", lags=24)(y=y_train, fh=3)

# Score forecasts in parallel
scores = mase(y_true=y_test, y_pred=y_pred, y_train=y_train)

# Forecast with target transforms and feature transforms
forecaster = linear_model(
    freq="1mo",
    lags=24,
    target_transform=scale(),
    feature_transform=add_fourier_terms(sp=12, K=6)
)

# Forecast with exogenous regressors!
# Just pass them into X
X = (
    y.select([entity_col, time_col])
    .pipe(add_fourier_terms(sp=12, K=6)).collect()
)
X_train, X_future = y.pipe(train_test_split(test_size=3))
forecaster = linear_model(freq="1mo", lags=24)
forecaster.fit(y=y_train, X=X_train)
y_pred = forecaster.predict(fh=3, X=X_future)

View the full walkthrough on forecasting here.

Feature Extraction

functime comes with over 100+ time-series feature extractors. Every feature is easily accessible via functime's custom ts (time-series) namespace, which works with any Polars Series or expression. To register the custom ts Polars namespace, you must first import functime in your module.

To register the custom ts Polars namespace, you must first import functime!

import polars as pl
import numpy as np
from functime.feature_extractors import FeatureExtractor, binned_entropy

# Load commodities price data
y = pl.read_parquet("https://github.com/functime-org/functime/raw/main/data/commodities.parquet")

# Get column names ("commodity_type", "time", "price")
entity_col, time_col, value_col = y.columns

# Extract a single feature from a single time-series
binned_entropy = binned_entropy(
    pl.Series(np.random.normal(0, 1, size=10)),
    bin_count=10
)

# 🔥 Also works on LazyFrames with query optimization
features = (
    pl.LazyFrame({
        "index": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
        "value": np.random.normal(0, 1, size=10)
    })
    .select(
        binned_entropy=pl.col("value").ts.binned_entropy(bin_count=10),
        lempel_ziv_complexity=pl.col("value").ts.lempel_ziv_complexity(threshold=3),
        longest_streak_above_mean=pl.col("value").ts.longest_streak_above_mean(),
    )
    .collect()
)

# 🚄 Extract features blazingly fast on many
# stacked time-series using `group_by`
features = (
    y.group_by(entity_col)
    .agg(
        binned_entropy=pl.col(value_col).ts.binned_entropy(bin_count=10),
        lempel_ziv_complexity=pl.col(value_col).ts.lempel_ziv_complexity(threshold=3),
        longest_streak_above_mean=pl.col(value_col).ts.longest_streak_above_mean(),
    )
)

# 🚄 Extract features blazingly fast on windows
# of many time-series using `group_by_dynamic`
features = (
    # Compute rolling features at yearly intervals
    y.group_by_dynamic(
        time_col,
        every="12mo",
        by=entity_col,
    )
    .agg(
        binned_entropy=pl.col(value_col).ts.binned_entropy(bin_count=10),
        lempel_ziv_complexity=pl.col(value_col).ts.lempel_ziv_complexity(threshold=3),
        longest_streak_above_mean=pl.col(value_col).ts.longest_streak_above_mean(),
    )
)

Related Projects

If you are interested in general data-science related plugins for Polars, you must check out polars-ds. polars-ds is a project created by one of functime's core maintainers and is the easiest way to extend your Polars pipelines with commonly used data-science operations made blazing fast with Rust!